https://github.com/jorisdg/SemanticKernel.PowerFx

Add low-code skills to Semantic Kernel using Power Fx
https://github.com/jorisdg/SemanticKernel.PowerFx

Last synced: 2 months ago
JSON representation

Add low-code skills to Semantic Kernel using Power Fx

• Host: GitHub
• URL: https://github.com/jorisdg/SemanticKernel.PowerFx
• Owner: jorisdg
• License: mit
• Created: 2023-08-21T18:54:15.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-11-30T19:52:58.000Z (about 1 year ago)
• Last Synced: 2024-08-01T13:19:28.341Z (5 months ago)
• Language: C#
• Size: 10.7 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-semantickernel - SemanticKernel.PowerFx - code skills to Semantic Kernel using Power Fx (Open Source Projects)

README

        
# SemanticKernel.PowerFx

This repo is a proof-of-concept for adding low-code native functions to Semantic Kernel using Power Fx.

## Semantic Kernel

[Semantic Kernel is an open source project](https://github.com/Microsoft/semantic-kernel) from Microsoft. It is an SDK that integrates Large Language Models (LLMs) like OpenAI, Azure OpenAI, and Hugging Face with conventional programming languages like C#, Python, and Java. Semantic Kernel achieves this by allowing you to define plugins that can be chained together in just a few lines of code.

## Orchestration

What makes Semantic Kernel (SK) most interesting is its orchestration and [planning capabilities](https://learn.microsoft.com/semantic-kernel/ai-orchestration/planner?tabs=Csharp). Given a set of plugins, SK is able to determine a list of plugins to use based on the _ask_ from a user. It is able to chain these together and pair inputs and outputs between them to come to an answer.

## Plugin functions

Beyond supporting the recently [standardized plugins for OpenAI](https://platform.openai.com/docs/plugins/getting-started/), Semantic Kernel (SK) allows you to create your own functions as plugins. SK supports **semantic** functions, which are basically prompts for an LLM with defined inputs and outputs, as well as **native** functions. [Native functions](https://learn.microsoft.com/semantic-kernel/ai-orchestration/native-functions?tabs=Csharp) are functions that are defined ahead of time in the host of the SK. These functions can be written in C# or Python.

## Low-Code Power Fx Functions

So the question is - why shouldn't we be able to define functions with low code as well? Suppose a low-code environment has the ability to expose its functions (predefined ones and/or user defined ones) to the Semantic Kernel, it would allow for users of that app to use natural language to execute commands. [Power Fx is another open source Microsoft project](https://github.com/Microsoft/Power-Fx) that provides a low code functional language with an Excel-like syntax.

## Disclaimer

The code here is provided as-is. It is my own personal project and just a proof-of-concept. This is not an officially sanctioned or supported project.